Academic publishing is undergoing radical transformation. Over the last decade, printed volumes have been increasingly supplemented or even replaced by digital editions due to changing customer demand, the requirements of funding organisations, and rising costs. The typically limited print-runs of many academic books have come up against the emerging ideal of open-access publications available to everyone, free of charge. On the other hand, as publishers have merged into ever larger conglomerates, scholar-led publishing in online repositories has begun to threaten their business model. In addition, image, video, and 3D content have conquered online spaces, and initiatives such as the European Open Science Cloud promote the idea of data spaces as the new methodological research paradigm. To address this new reality, including a push by funding organisations for open access in general and FAIR research data in standardised formats in particular, the Unit for Digital Resources of the Corpus Vitrearum (DU) has drafted initial recommendations on digital publishing. Based on an overview of past and current publishing endeavours along with best practices from the wider cultural-heritage domain, these recommendations aim to support national committees in their decision-making processes and to carry the successful tradition of a shared publishing guideline for Corpus Vitrearum volumes into the digital realm. Key questions that were addressed include: to what extent do digital-only publications need to deviate from the publication tradition outlined in the 2016 guidelines? What digital formats are best suited for which type of content? How can we simultaneously safeguard national data sovereignty and international long-term availability? How can existing print volumes best be retrodigitised to meet the demand for the digital availability of knowledge? In an age of data spaces, ‘generative AI’, and political threats to scientific data, how can we provide a digital equivalent to the trusted ‘blue volume’ identity of the Corpus Vitrearum without curtailing experimentation? The following aspects may serve as key takeaways. 1) Image or object catalogues provide highly structured data with the clear intention of supporting further research. As such, their exclusive availability in web applications is more likely to be accepted than that of, for example, essays. This type of synthesised research still gains prestige through book-like PDFs, traditional typesetting, and the name of a publisher or journal. 2) Across national projects, there have been numerous forays into digital publications, such as TEI XML files with web and print outputs or interactive hypertexts. Such endeavours are generally considered more complicated than established long-form text. 3) Digital workflows in traditional genres require further support, such as citation management or typesetting beyond Microsoft Word. 4) Born-digital publications seek to distinguish themselves from print publications by focusing on usability, data, quantification, video capture, digital twins, and other methods that, in turn, inform the type of synthesised research they provide. 5) Digital publications are more in tune with the FAIR principles, but the more extreme examples of ‘going digital’ also aspire to appeal to a wider audience than their more traditional counterparts.